Immune Particle filter Design for Terrain Referenced Navigation

Robust immune particle filter design for terrain‐referenced navigation with interferometric radar altimeter

“…To bring diversity to particle distribution, hyper‐mutation is executed. While previous hyper mutation process in [13] requires different kernel sizes for each particle, new hyper mutation process rescales a mutation kernel using tuning parameters, which significantly reduces computation time. Thus, the proposed IAPF is more robust and computationally effective compared to previous IAPF.…”

“…The previous IAPF [13] was originally applied in 1 mil IMU, but it tends to diverge on low roughness terrain with 100 mil IMU system. Thus, new IAPF is proposed to improve robustness on low roughness terrain.…”

“…The average computation time in each run is measured using MATLAB with Intel® Xenon CPU e5‐2620 v.4 @2.1 GHz, 64bit, 64 RAM. In this condition, APF takes 8 ms, previous IAPF [13] takes 30 ms and proposed IAPF takes 12 ms, respectively. The simulation results conclude that the proposed IAPF improves robustness of APF and benefits from computation cost.…”

“…The purpose of an immune system in an APF is to obtain optimal variance. Previous IAPF has been shown to enhance APF robustness [13] but it has required the presence of some inferior particles in order to maintain particle diversity. This indicates that the immune system has not managed to fully separate superior and inferior particles and due to leftover, computation cost is raised.…”